Clearly different durations and annotation rates for the same mgf file - why? #245
Comments
fsatfu
added
the
probable bug
|
Hi, This basically can not happen, or at least I don't know of any way it could, except for if the parameter sets were not actually identical Would your be able to share the two .Sirius projects with me? I might be able to understand what happened Please also share the input mgf and parameters used |
|
Hi, I just want to know, wether the data transfer worked? I did it last week? Or is there any mail adress I could send the link to the cloud to? (Last week I used Swisstransfer) Kind regards |
|
Hi, |
Hi, please provide the exact parameter set you used to compute these, else I cannot try to reproduce it |
|
Here is what I think happened: I'm looking at feature "77", m/z=143.990, RT 0.44. In your first project, this is the formula annotation, which leads to a top 1 hit in one of your databases
In your second project, this is the formula annotation (leads to no hit)
I can reproduce the results of the first project, if I compute the feature with these settings (Note that formula generation is set to DATABASE SEARCH)
And then I can also reproduce the result of the second project with these settings (Note that formula generation is set to DE NOVO + BOTTOM UP):
I'm pretty sure you forgot to change the formula generation setting for project 2 and thats why you are getting different results. To prevent this from happening, save your settings as a preset and just load it next time you are repeating the same analysis on different data |
|
Hi, |
|
Hi, sorry that I have to contact you again (as I said in the last 4 years everything allways worked well....) I repeated everything exectly with the same parameters I used in january this year with my first run and I also (settings as preset). The job stopped last night after five hours. I also used the same MZmine mgf file as the first time in january. Apr. 23, 2025 3:18:08 AM de.unijena.bioinf.jjobs.JJob lambda$logWarn$5 Apr. 23, 2025 3:18:08 AM de.unijena.bioinf.jjobs.JJob lambda$logError$2 Apr. 23, 2025 3:18:08 AM de.unijena.bioinf.jjobs.JJob lambda$logWarn$7 Apr. 23, 2025 3:18:08 AM de.unijena.bioinf.jjobs.JJob lambda$logWarn$5 Apr. 23, 2025 3:18:08 AM de.unijena.bioinf.jjobs.JJob lambda$logWarn$7 Apr. 23, 2025 3:18:08 AM de.unijena.bioinf.jjobs.JJob lambda$logWarn$5 Apr. 23, 2025 3:18:08 AM de.unijena.bioinf.jjobs.JJob lambda$logWarn$5 Apr. 23, 2025 3:18:08 AM de.unijena.bioinf.jjobs.JJob lambda$logError$2 Apr. 23, 2025 3:18:08 AM de.unijena.bioinf.jjobs.JJob lambda$logWarn$7 Apr. 23, 2025 3:18:08 AM de.unijena.bioinf.jjobs.JJob lambda$logWarn$7 Apr. 23, 2025 3:18:08 AM de.unijena.bioinf.jjobs.JJob lambda$logWarn$5 Apr. 23, 2025 3:19:38 AM org.springframework.web.servlet.mvc.support.DefaultHandlerExceptionResolver handleErrorResponse Apr. 23, 2025 6:52:05 AM org.dizitart.no2.mvstore.NitriteMVStore openMVMap Apr. 23, 2025 6:52:05 AM org.eclipse.jetty.ee10.servlet.ServletChannel handleException Apr. 23, 2025 6:54:38 AM org.springframework.web.servlet.handler.AbstractHandlerExceptionResolver logException |
Uh oh!
There was an error while loading. Please reload this page.
Bug description
We have recorded MS reference spectra for NP's from cyanobacteria for a larger project. We now wanted to do a kind of “before and after” study. For this, we used spectra from the MassBank of Europe as a user database (without our new data), then the smiles from the current CyanoMetDB for the first round of annotation. I did the first run at the end of January/beginning of February. (mgf file with MZmine version 4.4.3). Of the 3109 features, 886 were annotated. For 172 of these (11 directly NP's from Cyanos, 144 other compounds), we had matches with reference spectra from the MassBank EU and 239 annotations via our smiles. The runtime was almost four days at that time because we have around 260 compounds with over 800 Da. Now I wanted to repeat this process with exactly the same file (but as a new project) and I used exactly the same parameters. The first thing I noticed was that the calculation time was only just under 6 hours and that I only had 261 annotations in total (not 886 anymore), of which there were also significantly fewer matches with reference spectra. Also, only 97 annotations via our smiles database (239 before!). I have never repeated the same thing directly with a time delay, but somehow I can't reproduce the results from the first run. (I have used your great tool very often over the last almost four years and have never observed this before.) The only difference between the first run and the second was that I added the new reference spectra as .txt files as an additional customer database, just like the others and according to your instructions. So what could be the problem?
Expected behavior
I expected that with the addition of the new reference spectra we had recorded, I would receive significantly more annotations (library matches with reference spectra). I can see from the fact that some annotations were new and that I had matches with some reference spectra from the new database when I reviewed the results that our txt files must work. It also doesn't look like that the 800+ Da compounds would be missing?.
Steps To Reproduce
There was no error message. Only the clear differences in the results and the duration of the run. The mgf file was the same, so I don't have a different feature list. I also selected the same parameters under Compute all: same instrument (Orbitrap), same mass accuracy, same fallback adducts, molecular formula generation (de novo + bottom up), element filter as given, etc.
Screenshots and Log Files
Example data
System Info (please complete the following information):
Workarounds
No, but anyway thanks in advance for your help..... (mail adress: franziska.schanbacher@fu-berlin.de)
The text was updated successfully, but these errors were encountered: